JP4884789B2 - Ghost reduction in holographic stereograms - Google Patents

Description

  The present invention relates to a holographic stereogram. In particular, the present invention relates to ghosting reduction in holographic stereograms.

  A holographic stereogram is a type of hologram that is constructed or synthesized from a set of planar images of a subject that are continuously recorded on a holographic recording medium as a holographic element (hereinafter referred to as a hologel).

  FIG. 1 shows processing for generating a stereogram 10 with only horizontal parallax. Source data 12 in the form of a plurality of source data frames 14 including a planar image of object 15 is input to a processor 16 that divides each source data frame 14. Each section 18 of the source data frame 14 is then arranged using a remapping algorithm (eg, re-centering, panning or rotation, which indicates how the subject was captured by the video camera) The pre-distortion image 20 is generated, and the plurality of pre-distortion images 20 are, for example, spatial light modulators such as a liquid crystal display (hereinafter referred to as LCD). : Hereinafter referred to as SLM). The predistortion image 20 is projected onto a holographic recording medium 26 such as a photopolymer through a converging lens 24 (which is often a very complex lens assembly).

  The converging lens 24 compresses the predistortion image 20 into a plurality of holographic elements 28, each of which is continuously recorded on the holographic recording medium 26 to generate the holographic stereogram 10. .

  FIG. 2 shows a corresponding set of images 30 reproduced from the set of source data 12 and the holographic stereogram 10 generated by the technique shown in FIG. This roughly shows an image that the observer actually sees. Lightness, contrast, etc. may be different. Ideally, the holographic stereogram 10 accurately reproduces the source data 12. However, as shown in FIG. 2, a shadow often appears in the image 30 reproduced from the holographic stereogram 10. The “shadow” that appears in the image 30 reproduced from the holographic stereogram 10 is called a ghost. Due to this ghost phenomenon, the fidelity of the image 30 of the holographic stereogram is impaired.

  Thus, in view of the above-mentioned problems, the object of the present invention is to reduce ghosts in an image reproduced from a holographic stereogram so that the final stereogram image reproduces the original image 30 as accurately as possible. Or generating a holographic stereogram to remove.

  The present invention provides an image processing method for generating a holographic stereogram and a stereogram generated thereby. The present invention can be realized in various forms, for example, as a process, an apparatus, a system, a device, or a method.

  As a first aspect of the present invention, the present invention provides an image processing method for generating a holographic stereogram from a plurality of source data frames. A plurality of source data frames are divided into a plurality of sections, and the plurality of sections are mapped to generate a plurality of predistortion images. Compensates for variations in saturation of multiple predistortion images, compresses multiple predistortion images, generates multiple holographic elements, records multiple holographic elements on a holographic recording medium, and generates holographic stereograms Generate.

  Preferably, the plurality of predistortion images are compensated by applying a compensation algorithm to one or more pixels in the plurality of predistortion images. The compensation algorithm adjusts one or more pixels in the plurality of predistortion images, for example, when a plurality of holographic elements are recorded on the holographic recording medium, and at least partially compensates for subsequent saturation variations. . One or more attenuation factors may be applied to one or more pixels of a plurality of predistortion images. For example, the pixel value may be reduced with respect to pixel values along the same compression axis in the same frame or other frames, depending on the attenuation rate applied to the pixel. The attenuation rate applied to the pixel may reduce the pixel value to the lowest pixel value level along the same compression axis.

  Preferably, the image processing method includes the step of establishing a graphical relationship between the average pixel value of the plurality of pixels of the plurality of predistortion images and the brightness of the image reproduced from the holographic stereogram. The average pixel value can be averaged over the compression axes of a plurality of predistortion images. This graphical relationship includes the steps of generating a series of reference source data frames with color gradients, dividing and mapping the series of reference source data frames to generate a set of predistortion images, and a set of predistortion images. Compressing the distortion image to generate a set of holographic elements; recording the set of holographic elements to a second holographic recording medium to generate a second holographic stereogram; and Measuring the brightness of the image reproduced from the two holographic stereograms, and the average pixel value of the pixels in the set of predistortion images for the brightness of the image reproduced from the second holographic stereogram And plotting a graph of It may be.

  Based on the average pixel value measured from the predistortion image, using a graphical relationship between the average pixel value of multiple predistortion inspection images and the brightness of the image reproduced from the holographic stereogram The brightness of the actual image reproduced from the final holographic stereogram can be predicted. The predistortion image can be compensated based on the predicted brightness.

  A weighting factor may be applied to the predicted brightness based on the color. Such a weighting coefficient can be obtained from a graph of the weighting coefficient of one color component of the pixel with respect to the average pixel value of the other color components of the pixel. The graph preferably includes a first source data frame and a second source having a plurality of sections each having a different combination of color components, the second source data frame having a black region on one side. Generating a data frame, deriving a first image and a second image from the first source data frame and the second source data frame, respectively, calculating a reference ratio from a reference section of the second image; Calculate the ratio of each remaining section of the image, divide the ratio of each remaining section by the reference ratio to obtain coordinates along the weighting factor axis, one coordinate represented along the weighting factor axis, and the other Are plotted by plotting a plurality of points represented along the respective average pixel value axes of other color components in the pixel, and creating a graph.

  As a second aspect of the present invention, the present invention provides an image processing method for reducing ghosts in a holographic stereogram. The plurality of source data frames are divided, and the divided sections are mapped to generate a plurality of predistortion images, and compensate for variations in saturation of the plurality of predistortion images.

  As a third aspect of the present invention, the present invention provides an image processing method for modifying a predistortion image and generating a holographic stereogram from a plurality of source data frames. The predistortion image is generated by dividing the source data frame into a plurality of sections and mapping the plurality of sections, and compensates for variations in saturation of the plurality of predistortion images.

  In the second and third aspects, a plurality of source data frames are divided into a plurality of sections before the plurality of predistortion images are compensated, and the sections are mapped to generate a plurality of predistortion images. In the third and fourth aspects, after compensating a plurality of predistortion images, the compensated predistortion images are compressed, a plurality of holographic elements are generated, recorded on a holographic recording medium, and a holographic stereogram Is generated.

  As a fourth aspect of the present invention, the present invention provides a holographic stereogram generated based on any of the first to third aspects.

  As a fifth aspect of the present invention, the present invention provides an image processing apparatus for generating a holographic stereogram from a plurality of source data frames. An image processing apparatus includes: a dividing unit that divides a plurality of source data frames into a plurality of sections; a mapping unit that maps a plurality of sections to generate a plurality of predistortion images; and at least one of the plurality of predistortion images Compensation means for compensating a plurality of predistortion images by applying a compensation algorithm to the pixels.

  As a sixth aspect of the present invention, the present invention provides a computer system for generating a holographic stereogram from a plurality of source data frames. The computer system includes an external device interface that receives a plurality of source data frames. A processor connected to the external device interface divides and maps a plurality of source data frames to generate a plurality of predistortion images. A memory connected to the processor is applied to one or more pixels and stores a compensation algorithm that compensates for saturation variations in multiple predistortion images.

  As a seventh aspect of the present invention, the present invention provides a computer program product for generating a holographic stereogram from a plurality of source data frames. The computer program product divides and maps a plurality of source data frames connected to an external device interface to generate a plurality of predistortion images, a computer readable program code, and saturation variations in the plurality of predistortion images. And a computer readable program code for applying a compensation algorithm to one or more pixels.

  The image processing apparatus according to the fifth aspect, the computer system according to the sixth aspect, and the computer program product according to the seventh aspect may all operate based on the image processing method according to the first aspect to the third aspect. Good.

  Other aspects and advantages of the invention will become apparent from the following detailed description, taken in conjunction with the drawings, illustrating by way of example the principles of the invention.

  A new technique for generating holographic stereograms is disclosed. In the following, numerous specific details are set forth in detail in order to clarify the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced without some or all of these specific details. In other instances, well known process operations have not been described in detail in order not to unnecessarily obscure the details of the embodiments of the present invention.

  The inventors have studied the ghost phenomenon of holographic stereograms and found that the ghost phenomenon occurs when a holographic element is recorded on a holographic recording medium. That is, it has been found that ghosts are due to the non-linear characteristics of the recording medium and the resulting variation in saturation level. This gives the impression that the holographic element is “contaminated” by data from adjacent source data frames.

  The mechanism for generating ghosts is due to limitations in the recording material (eg, photopolymer). The diffraction efficiency (brightness) of a photopolymer does not increase linearly or indefinitely with increasing laser power applied to the photopolymer. When the laser power applied to the hologel is small (determined by the predistortion image), the performance of the photopolymer is “not strained”. On the other hand, when the laser power applied to the hologel is large (determined by a predistortion image with a high palette value), the performance of the photopolymer is “strained” and quickly reaches the performance limit, ie Saturates.

  FIGS. 3A to 3C show changes that occur when a series of predistortion images are recorded as holographic elements on a holographic recording medium, based on conventional technology processing. Specifically, FIG. 3 (a) shows a series of predistortion images, and FIG. 3 (b) shows a series generated from the series of predistortion images of FIG. 3 (a) recorded on a holographic recording medium. FIG. 3 (c) shows the series of holographic elements of FIG. 3 (b) extended to represent a predistortion image.

  Saturation (i.e., increasing the laser power applied to the recording medium, the brightness gradually decreases or flattens) is observed in holographic elements recorded on the holographic recording medium. However, due to the appearance of ghosts, the degree of saturation of the pixels constituting the holographic element varies. In other words, the ghost phenomenon indicates the saturation of several pixels that make up the holographic element. As a result, in the holographic element, a plurality of tones 32, 34, 36, and 38 of a single color 40 are observed in the predistortion image.

  In order to reduce the ghost phenomenon, we apply to the predistortion image, and in the predistortion image, by adjusting the related pixel value by the related amount, the pixel saturation variation resulting from the ghost phenomenon. We devised a compensation algorithm to compensate. This algorithm predicts saturation and reduces specific pixel values in later processing. The compensation algorithm can be written into software that is processed by a processor that generates a predistortion image based on the source data.

  FIG. 4A shows a predistortion image before applying a compensation algorithm according to an embodiment of the present invention. The predistortion image can be generated from the source data as in the conventional technique.

  In this embodiment, four source data frames are used, but the number of source data frames to be used may be more or less than four. Usually, the number of such frames is in the range of 150-225. Similarly, in this example, only four predistortion frames (out of a total of 50) are shown, but more frames may be used. The number of predistortion frames in the compression axis is the number of hologels in the same compression axis in the generated holographic stereogram. For example, if the width of the final holographic stereogram is 45 mm and the width of each hologel is 0.2 mm, the number of hologels is 225, so 225 predistortion images are required. Further, when the width of the final holographic stereogram is 10 mm and the width of each hologel is 1.0 mm, the number of hologels is 10, and thus 10 predistortion images are required.

  The source data can be any data including, for example, a bit-mapped graphic (BMP), a joint photographic expert group (JPEG), or an audio video interleaved (AVI). It may have a format. Correspondingly, each source data frame may be an independent bitmap image in a continuous BMP stream, a single JPEG image file, a still image in an AVI file, or the like.

  FIG. 4 (b) shows a compensated predistortion image according to one embodiment of the present invention. In FIG. 4B, due to the ghost phenomenon, the pixels in the upper region and the lower region having the highest average pixel value are more saturated than the pixels in the central regions 42, 44, 46. Here, if no adjustment is made, the central regions 42, 44, 46 appear brighter than the upper and lower regions. For this reason, the pixel values of the central regions 42, 44, 46 that are not saturated as much as the upper region and the lower region are reduced. Thus, the central region coincides with the upper region and the lower region, and the pixel value of any color (in this case, green) is the same in the final holographic stereogram (eg, R0, G255, B0).

  FIG. 4 (c) shows an uncompressed holographic element generated from the compensated predistortion image according to one embodiment of the invention shown in FIG. 4 (b). By compensating for the saturation of some of the pixels that make up the holographic element, a single color is observed around the portion corresponding to the image object in the uncompressed holographic element.

  When recording holographic elements generated from a predistortion image on a holographic recording medium, the inventors adjust the pixels affected by the predistortion image so as to compensate for the later reduced saturation. The following compensation formula (1) was derived.

Where palette _New represents the compensated pixel value of the pixel in the predistortion image, palette represents the original pixel value of the pixel in the predistortion image, and atten is a palette to reduce ghosting. It represents the attenuation rate that determines the degree of adjustment.

  Each pixel can be represented by three component values: a red component, a green component, and a blue component. Each component value indicates the intensity of a specific color component of the pixel. Therefore, the pixel value including the value of each color component indicates the color of the pixel.

  The number of unique colors that can be represented by a pixel depends on the number of bits per pixel (bpp). In the 24 bpp pixel format, each of the red / green / blue (RGB) components is composed of 8-bit bytes and represents a numerical value in the range of 0-255. Therefore, a pixel in 24 bpp mode can represent 16.7777216 colors. For example, a pixel having a value of (0, 0, 0) meaning (red = 0, green = 0, blue = 0) is black, and a pixel having a value of (255, 255, 255) is white. , Pixels having a value of (255, 0, 0) are red, pixels having a value of (255, 255, 0) are yellow, and pixels having a value of (255, 165, 0) are orange is there.

  In order to explain the present invention, a 24-bpp mode is employed in the embodiment of the present invention. Note that the present invention may be applied to any other mode including, for example, an 8 bpp pixel format in which pixels can have 256 colors or a 16 bpp pixel format in which pixels can have 65,536 colors. It is clear.

  The purpose of the attenuation factor atten is to reduce the actual pixel value palette of the pixel in the predistortion image to the lowest pixel value (excluding the pixel representing the object) in the holographic element recorded on the holographic recording medium. That is.

  Here, it is assumed that the source image has only one tone of one color (the object itself is ignored). If there is only one shade of the original source image, there should be only one shade of the final hologel print. Therefore, multiple tones should be excluded (so that only one tone exists). This process can simply be simplified to two steps.

  Step 1: Identify the darkest tone.

  Step 2: Reduce the brightness of all lighter shades to match the darkest shade.

  The central hologel (stretched) shown in FIG. 3 (c) shows four different shades: the darkest shade 32 in the upper and lower strips of the image and the brightest in the center. The second lightest region shows the three lighter banded regions 34, 36, 38 adjacent to the darkest shade 32. The aim here is to reduce the brightness of the three lighter bands in the printed hologel and to match all these shades to the darkest shade 32 target shade. A desirable constant color tone is shown in FIG.

  The lowest pixel value of the holographic element depends on the performance of the holographic recording medium. When the intensity of light applied to the holographic recording medium is constant, the higher the performance of the holographic recording medium, the brighter the holographic element formed. The performance of the recording material is affected by factors such as the laser power to be exposed.

  The high performance of the recording material can be easily confirmed by holographic stereogram printing. That is, the brighter the holographic image, the higher the performance of the recording material. This performance is generally called diffraction efficiency (hereinafter referred to as DE) of the recording material. Simply put, the higher the DE, the more light is reflected and the image appears brighter to the viewer. Once the holographic stereogram is recorded, the diffraction efficiency of each part of the recording material is different.

  The attenuation rate atten is first determined from the relationship between the average pixel value pAvg along the compression axis of the predistortion image and the brightness Br of the corresponding pixel in the image reproduced from the uncompensated holographic stereogram. This relationship is preferably determined by the same holographic recording material used to record the target hologram. Once the relationship between the average pixel value pAvg and the brightness Br is established, the brightness of the individual pixels making up the uncompensated holographic element is the average pixel value of the pixels making up the corresponding predistortion image. It can be predicted from pAvg. Thereby, when the holographic element is recorded on the holographic recording medium, the pixel affected by the ghost phenomenon can be adjusted by compensating the inaccurate brightness to an appropriate brightness.

  FIG. 5 shows a part of the technique for establishing the relationship between the average pixel value pAvg and the brightness Br for one of the RGB components. First, as shown in FIG. 5, a series of reference data frames or source data frames 50 is generated so that the number of frames covered by the object gradually increases. Next, as in the prior art, each source data frame 50 is divided and arranged by a mapping algorithm to generate a set of predistortion images. In addition, the predistortion image is compressed to produce a set of holographic elements 52. Then, the holographic element 52 is continuously recorded on the holographic recording medium, and the holographic stereogram 54 is formed by a conventional method. For clarity of explanation, FIG. 5 shows the holographic element 52 in an uncompressed form.

  Here, with respect to the holographic stereogram 54, a charge coupled device (CCD) or a spectroscope is arranged at a plurality of viewpoints 56, and a series of images 58 corresponding to the source data frame 50 reproduced from the holographic stereogram 54. The brightness Br is measured.

  For the same frame used to create the holographic stereogram 54, the value of the average pixel value pAvg is determined.

  In this example, there are 10 different frames and 10 hologels. The 10 frames show 10 different shades of a single color (eg green) except for the objects in the last hologel, which are in a predistortion image (these images are all the same) Resulting from 10 different average pixel values pAvg values. These different shades appear in 10 different layers. By determining the brightness value of the holographic stereogram at locations corresponding to 10 different layers, a curve linking the brightness Br to the average pixel value pAvg can be plotted.

  Each predistortion image has n rows and i columns of pixels. The average pixel value pAvg of the pixels 62 in each row n of the pixels in the predistortion image 64 is calculated based on the following equation.

Here, num_horiz_pix represents the total number of pixels across the compression axis, and pAvg _n represents the pixel value averaged over the compression axis of row n of pixels.

  In the case of a color stereogram, the average pixel value pAvg of each RGB component is calculated separately.

  Each pixel 62 of the predistortion image 64 has a minimum pixel value of 0 in the case of black, and a maximum pixel value of 255 in the case of a specific primary color (for example, green). The average pixel value pAvg is calculated by calculating the sum of all the pixel values of the pixels 62 along the compression axis and dividing the sum by the total number of pixels along the compression axis. Although a horizontal parallax system is described here, it is obvious that the present invention can be applied to a vertical parallax system or an omnidirectional parallax system. The omnidirectional parallax system will be described later.

  As shown in FIG. 6, there may be hundreds of pixels in each of the horizontal and vertical directions, but the fact that there are only 10 frames, the 10 different predistortion images across the compression axis. It means that it can be divided into strips. For all rows, the pixel values in each strip are all the same. Furthermore, because of the nature of these specific source data frames, the predistortion image can be considered to have 11 different layers. Within each layer, the pixel values are all the same for any column of pixels, and within each layer, the rows of pixels are the same. Since there are only 11 such layers and they are all different, 11 different average pixel values pAvg are obtained as shown in Table 1 below. In Table 1, this example assumes a monochrome image that is green and shows only one color channel. When there are two or more colors, the same calculation is performed for each color.

  Here, for the sake of brevity, only one value is shown for each strip in each layer, rather than performing an operation on each row.

  The brightness Br of the pixel of the image 58 derived as described with reference to FIG. 5 and the average pixel value pAvg of the corresponding pixel 62 of the predistortion image 64 derived as described with reference to FIG. Provide a set of points. The brightness Br of each different viewpoint of the image 58 and the average pixel value pAvg of the corresponding pixel layer in the predistortion image constitute a point on the graph of the average pixel value pAvg versus the brightness Br. Using the set of these points, a graph of average pixel value pAvg versus brightness Br is plotted. Rows (and layers) with an average pixel value pAvg of 0 are ignored. As a result, the leftmost viewpoint 56 of the holographic stereogram 54 corresponds to the tenth layer (first layer that is not 0) of the predistortion image 64. The predistortion image 52 is an implicit representation of the original image 50. When the eye is in the leftmost viewpoint 56, the “leftmost” column of each holographic element 52 is visible to that eye. All these visible “assemblies” of the columns together regenerate the leftmost viewpoint 56. The rightmost viewpoint 56 of the holographic stereogram 54 corresponds to the first layer of the predistortion image 64. When the eye is in the rightmost viewpoint 56, the eye sees the “rightmost” row of holographic elements 52. All these visible “assemblies” of the columns together regenerate the rightmost viewpoint 56.

  For example, as will be described later with reference to FIG. 8A, an optimal nth-order polynomial indicating the relationship between the average pixel value pAvg and the brightness Br at various points (10 points in this specific example). (For example, a quintic or sixth-order polynomial function representing a curve as accurately as possible).

  5 and 6, the case of a single color was examined. However, images often have different colors. For this reason, such measured values and graphs of the reference image are created for each of red, green and blue on the recording material.

  In FIG. 5, the viewing angle covering all viewpoints is about 90 °. In the case of 10 frames, different images are generated by setting each interval 60 of the viewpoint 56 to 9 °. When it is determined that the interval 60 between successive viewpoints 56 is too narrow to obtain a substantially distinguishable measurement value of brightness Br (this interval may be 9 °), FIG. 7 is used. As described above, the brightness Br may be obtained using a slightly different method.

  In the example shown in FIG. 7, a series 70, 72 of first and second reference data frames or source data frames having a color gradation is generated. In practice, the first and second source data frame series 70 and 72 shown in FIG. 7 are created from the same source data frame as in FIG. However, the first source data frame series 70 is configured such that every other frame in the source data frame series 50 shown in FIG. 5 is repeated once (that is, appears twice). . Similarly, as shown in FIG. 7, the second source data frame series 72 is configured to repeat every other frame in the source data frame series 50 of FIG. 5 one at a time. . The first and second sets of predistortion images are then generated by dividing and arranging these first and second source data frame series 70, 72 via a mapping algorithm. The first and second sets of predistortion images are compressed into first and second sets of holographic elements, respectively, which are continuously recorded on the first and second holographic recording media, First and second holographic stereograms 74, 76 are formed. Here, for each holographic stereogram 74, 76, a charge coupled device (CCD) or a spectrometer is arranged at a plurality of viewpoints 78, and the source data frame 70, reproduced from the holographic stereogram 74, 76, The brightness Br of a series of images 80 corresponding to 72 is measured.

  The brightness Br of pixel of image 80 and the average pixel value pAvg of the corresponding pixels in the two corresponding predistortion images plot a graph of average pixel value pAvg vs. brightness Br and set a set of points to draw a curve. provide.

  If the spacing 82 between each successive viewpoint 78 is still too narrow to obtain a substantially distinguishable measurement of brightness Br, the process shown in FIG. 5 to increase the number of repetitions of the source data frame 50 in FIG. 5 until the spacing between successive viewpoints is sufficiently large to obtain a distinctly distinct brightness Br measurement. 50 may be generated.

  FIG. 8A shows the relationship between the brightness Br generated from the reference image and the average pixel value pAvg for one color component of one recording material. As long as all values are measured in the same way, the unit of brightness Br is arbitrary or irrelevant. When drawing a curve, the brightness Br corresponding to the first layer (having the lowest brightness Br and the highest average pixel value pAvg) is given a value of 1.0 (which compensates for the predistortion image). In this case, the attenuation factor atten of other rows / layers having the same average pixel value pAvg is 1, ie, such rows / layers are no longer darkened), correspondingly to the first layer The other brightness Br plotted by shifting by the same amount as the brightness Br corresponding to is shifted so as to give a value of 1.0.

As a modification, the average pixel value pAvg obtained by the following equation (3) may be expressed as a ratio pAvg _ratio .

Here, palette _Max represents the maximum pixel value based on the color mode being used, and is 255 in this specific example.

Accordingly, the curve shown in FIG. 8B in which a set of points of lightness Br with respect to the average pixel value pAvg _ratio is plotted shows the relationship between the average pixel value pAvg _ratio and the lightness Br for one of the RGB components. Again, the minimum value of brightness Br is set to 1.0.

  As mentioned above, the ghost phenomenon reduces the saturation of the pixels in the holographic elements that make up the holographic stereogram. As a result, the attenuation rate atten is inversely proportional to the brightness Br of the pixels in the image reproduced from the uncompensated holographic stereogram. In this specific example, as will be described below, the derived attenuation rate atten is the reciprocal of the brightness Br.

  FIG. 8C shows this relationship.

Once the relationship between the brightness Br and the average pixel value pAvg or the relationship between the average pixel value pAvg _ratio and the brightness Br is determined from the reference image, the average from the actual image converted to a holographic stereogram The measured value of the pixel value pAvg can be used to derive the corresponding brightness Br for realizing correct compensation.

  FIG. 9 shows a method for calculating the average pixel value pAvg of the predistortion image 48 derived from the source data frame reduced to a holographic stereogram. Each pixel 46 of the predistortion image 48 is 0 when the minimum pixel value is black, and the maximum pixel value is 255. The average pixel value pAvg is obtained by calculating the sum of the pixel values of all the pixels along the compression axis based on the equation (2), and this sum is calculated as the total number of pixels along the compression axis (that is, the width of the pixel unit for the predistortion image ).

  The average pixel value pAvg needs to be calculated for all pixel rows. Therefore, when the image height is 480 pixels, it is theoretically necessary to calculate the average pixel value pAvg for all 480 pixel rows. Here, for the sake of brevity, only four frames are shown, so only seven measurement values calculated using the four pixel values are shown. In FIG. 9, square brackets are shown beside the region where the average pixel value pAvg is constant.

  Although a horizontal parallax system is described here, it is obvious that the present invention can be applied to a vertical parallax system or an omnidirectional parallax system. The omnidirectional parallax system will be described later. In the vertical parallax system, various processes applied to the horizontal direction and the vertical direction described above are reversed and applied to the vertical direction and the horizontal direction.

  The predistortion image 48 shown in FIG. 9 has four frames and seven distinguishable layers 46. This potentially yields seven different average pixel values pAvg, but in practice only four different average pixel values pAvg are used, as shown in Table 2 below.

The average pixel value pAvg can be measured for all the pixel layers of the predistortion image. From the graph of the average pixel value pAvg versus brightness Br shown in FIG. 8A (or using the pAvg _ratio from the graph of average pixel value pAvg _ratio vs. brightness Br shown in FIG. 8B), each average pixel value pAvg A predicted value of brightness Br can be obtained for. The attenuation rate atten may be calculated for each pixel of the predistortion image by substituting the predicted value of the brightness Br into Expression (4). Instead of this, the expression (4) may be expressed as a graph shown in FIG. 8C with respect to the range of the brightness Br plotted in the graphs of FIG. 8A and FIG. 8B. The predicted attenuation rate atten of the brightness Br of the image may be read from the graph of Br vs. atten shown in FIG. As a further variation, a single graph of atten vs. pAvg (or atten vs. pAvg _ratio ) may be plotted.

Based on the graphs of FIGS. 8A to 8C and Expression (1) “palette _New = palette × atten” and Expression (4), two of the three color components of the pixel have 0 value and palette _New . The compensated pixel value of the pixel of the predistortion image in the monochrome system can be calculated by any one of the following equations (5), (6), and (7).

Color System In a color system in which a holographic stereogram includes two or more color components of the three RGB color components, the predicted brightness Br of a single RGB component affects the pixel values of other RGB components. Receive. Since the pixel value of a pixel is not proportional to the sum of the pixel values of each of the three RGB components, the embodiment for the color system uses a weighting factor Weight _X to determine the pixel values for a single RGB component. Compensates for the effects of pixel values of other RGB components. That is, the brightness Br can be expressed by the following equation (8).

Here, Br _{X (color)} represents the lightness of the component X (one of RGB components) of the pixel in the color system, Br _{X (mono)} represents the lightness of the X component of the pixel in the monochrome system, and Weight _X represents the action of the other two RGB components Y and Z on Br _{X (mono)} .

Weight _X is a function of the average pixel values pAvg _{ratio, Y} and pAvg _{ratio, Z} of each of the other two RGB components of the pixel of the predistortion image.

Similar to the technique described above for the monochrome system, a graph of the average pixel value pAvg _ratio vs. brightness Br is plotted, and the curve drawn for the different color combinations is one average pixel value pAvg _ratio for the RGB component for each combination. And the lightness Br. FIG. 10 shows three curves of average pixel value pAvg _{ratio, G} vs. Br _G. The uppermost curve is a curve with R = B = 0, which is the same as the curve in FIG. The middle curve represents the intermediate value of R and B where R and B are not zero. The bottom curve is a curve with R = 255 and B = 255. It can be seen from these sets of curves that the weight curves of different colors are actually surface curves as shown in FIGS. 11 (a), 11 (b) and 11 (c).

Therefore, the relationship between the weighting factor Weight _X and the average pixel values pAvg _{ratio, Y} and pAvg _{ratio, Z} is a graph as shown in FIG. 11 (a), FIG. 11 (b) and FIG. 11 (c). Can be expressed. FIG. 11A is a graph showing the relationship between the weighting coefficient Weight _G and the average pixel values pAvg _{ratio, R} and pAvg _{ratio, B.} FIG. 11B is a graph showing the relationship between the weighting coefficient Weight _R and the average pixel values pAvg _{ratio, G} and pAvg _{ratio, B.} FIG. 11C is a graph showing the relationship between the weighting coefficient Weight _B and the average pixel values pAvg _{ratio, R} and pAvg _{ratio, G.}

  Hereinafter, a method for creating the graph of FIG. 11A based on an embodiment of the present invention will be described with reference to FIGS. 12A, 12B, and 13. FIG.

  In this method, a set of source data frames 200 is generated as shown in FIG. 12A, and as a result, a set of images 212 is generated as shown in FIG. 12B.

  The set of source data frames 200 includes a first source data frame 202 and a second source data frame 204. The first source data frame 202 includes a first set of sections 206, each section 206 having a different combination of RGB components and thus a different color. Table 3 below shows the pixel values of each RGB component of each section 206 of the first source data frame 202.

In this specific example, the pixel value of the green component of the first source data frame 202 is kept constant at the maximum pixel value, and the effect of the presence of the blue and / or red component on Br _G is determined. Different colors for each section 206 are obtained by changing the pixel values of the blue and red components of each section 206. The effect of the presence of blue and / or red components on Br _G becomes more apparent by keeping the green component pixel value constant at the maximum pixel value.

  On the other hand, the second source data frame 204 includes a second section set 208 having the same color as the first section set 206. However, the pixel value of each RGB component in the portion 210 of the second source data frame 204 is zero, and a black region is added to the second source data frame 204 as shown on the right side of FIG. Form.

  Then, as described above, a first holographic stereogram is generated by printing the set 200 of source data frames on a holographic recording medium.

  FIG. 12 (b) shows a set of images 212 of a set of source data frames 200 reproduced from the first holographic stereogram. The set of images 212 includes a first image 214 and a second image 216. The first image 214 is a reproduction of the first source data frame 202, and the second image 216 is a reproduction of the second source data frame 204. Saturation is observed in the right portions 218, 220 of the first image 214 and the second image 216, respectively.

  FIG. 13 shows calculation of points for plotting the graph of FIG. 11A from the image 214 of FIG. The value from the image 214 can be measured using, for example, a commercially available CCD having a manual override function (for example, Sony Servershot DSC-V1 (trademark), etc.). The manual override function allows the user to set shooting conditions such as 1) focus distance, 2) white balance, 3) f-number, and 4) aperture.

  Since all of the operations performed are related only to the green component, the color of the first image 214 is set to three RGB, for example, commercially available image editing software (eg, Photoshop ™). Software that can be divided into respective pixel values of the components may be used to determine the pixel values of the green component in various sections of the first image 214 and the third image 264.

  A reference ratio ref is generated for the first image 214 in FIG. The reference ratio ref is the ratio of the pixel value of the saturated section to the unsaturated section, and the unsaturated section contains only green, i.e., R = 0, G = 255, B = 0 (first The uppermost layer of the image 214 of FIG.

In addition, for each of the remaining sections of the first image 214, ratios r ₁ , r ₂ ... R ₈ of pixel values of the saturated section to the non-saturated section are also calculated.

By dividing each of the ratios r ₁ , r ₂ ... R ₈ by the reference ratio ref, points pt ₁ , pt ₂ for plotting the graph of FIG. 11A along the weighting coefficient Weight _G axis. ... the coordinates of pt ₈ are obtained.

Each of the points pt ₁ , pt ₂ , pt ₃ , pt ₄ , pt ₅ , pt ₆ , pt ₇ , pt ₈ has coordinates along the Weight _G axis, coordinates along the pAvg _{ratio, R} axis, and pAvg _{ratio , And} coordinates along the _B axis, ie, (Weight _G , pAvg _{ratio, R} , pAvg _{ratio, B} ) and plotted as shown in FIG. The average pixel values pAvg _{ratio, R} and pAvg _{ratio, B-} axis coordinates are obtained from the red and blue component pixel values in the individual curves of section 206 of the first source data frame 202, respectively.

  Then, in order to create the graphs of FIGS. 11B and 11C, the same process as the process of creating the graph of FIG. 11A is repeated. 11B, the pixel value of the red component in the source data frame is kept constant, and the points in the graph of FIG. 11B are calculated based on the pixel value of the red component in the reproduced image. The For the graph of FIG. 11C, the pixel value of the blue component in the source data frame is kept constant, and the points of the graph of FIG. 11C are calculated based on the pixel value of the blue component in the reproduced image. The

As is apparent from the graph, when the pixel values of the other two components are 0, the weighting factor Weight _X is constant at a value of 1, and Br _{X (colour)} is equal to Br _{X (mono)} .

According to the graphs of FIGS. 11A to 11C, the compensated pixel value pallet _New of the RGB components of the pixels of the predistortion image of the color system is expressed by any one of the following formulas (13) to (21). Can be used to calculate.

For example, the compensated pixel value palette _New , _{G (colour)} of the pixel of the predistortion image pixel in the color system calculates the green component pixel value palette _G , determines the average pixel value pAvg of the pixel, and based on this By obtaining Br _{G (mono} ) from FIG. 8 (a), obtaining the average pixel values pAvg _{ratio, R} and pAvg _{ratio, B} of the pixels, and based on this, the weighting factor Weight _G is obtained from FIG. 11 (a). It can be calculated using equation (14).

  FIG. 14 is a flowchart illustrating a method 300 for generating a holographic stereogram according to an embodiment of the present invention. The method 300 includes inputting 302 source data including a plurality of source data frames including a two-dimensional image of a subject to a processor. In step 304, the processor divides the source data frame. In step 306, each section of the source data frame is mapped to generate a plurality of predistortion images. In step 308, the ghost phenomenon is reduced by applying a compensation algorithm to a plurality of pixels of the predistortion image to compensate for variations in saturation of the predistortion image. Thereafter, the compensated predistortion image is input to the SLM. In step 310, the compensated predistortion image is compressed to form a plurality of holographic elements, and in step 312, each of the plurality of holographic elements is continuously recorded on a holographic recording medium to obtain a holographic stereogram. Is generated.

  Details of step 308 for compensating each of the predistortion images are shown in FIG. Specifically, FIG. 15 is a flowchart illustrating a method 400 for compensating for variations in saturation of predistortion images, in accordance with one embodiment of the present invention. In the method 400, first, in step 402, it is determined whether the predistortion image is a monochrome image or a color image.

  If it is determined in step 402 that the predistortion image is a monochrome image, in step 404, the pixel value palette and average pixel value pAvg of each pixel of the predistortion image are determined.

  Next, in step 408, a graph of predicted brightness Br vs. average pixel value pAvg is created based on the method described above with reference to FIGS. Next, in step 410, the predicted brightness Br is determined based on the predicted brightness Br versus average pixel value pAvg graph.

In step 412, the compensated pixel value pallet _New of each pixel in the predistortion image is calculated using any one of the equations (5), (6), and (7).

On the other hand, if it is determined in step 402 that the predistortion image is a color image, in step 414, the respective pixel values palette _X and average pixel value pAvg _X of the RGB components of each pixel of the predistortion image are determined.

Next, in step 418, for example, based on the method described above with reference to FIGS. 5 to 9, assuming that each of the R, G, and B components is X, the predicted brightness Br _X versus the average pixel value pAvg _X Create a graph. Next, in step 420, each of the R, G, and B components is set as X, and the predicted brightness Br _X is determined based on the predicted brightness Br _X vs. average pixel value pAvg _X graph.

Next, in step 424, for example, based on the method described above with reference to FIGS. 11 to 13, each of the R, G, and B components is set to X, and the average pixel values pAvg _Y and pAvg _Z vs. the weighting coefficient Weight _X Create a graph. In step 426, the weighting coefficient Weight _X is determined from the graph of the average pixel values pAvg _Y and pAvg _Z versus the weighting coefficient Weight _X , with each of the R, G, and B components as X.

Next, in step 428, the compensated pixels of each RGB component of each pixel of the predistortion image based on the expressions (13) to (15), (16) to (18), and (19) to (21) The values palette _{New, R} and palette _{New, G} and palette _{New, B} are calculated.

  The method 400 shown in FIG. 15 is repeated for each predistortion image generated from the source data frame.

Omni-directional parallax The process and embodiments described above are for a horizontal parallax holographic stereogram that produces an image in only a single plane. The present invention can also be applied to an omnidirectional parallax holographic stereogram that generates an image on a plurality of planes.

  The generation of the omnidirectional parallax holographic stereogram according to the present invention uses many aspects used for the horizontal parallax holographic stereogram described above.

  A set of source data frames 502 for an omnidirectional parallax holographic stereogram is shown in FIG. The source data frame 502 includes a plurality of series of frames 504a, 504b, 504c, 504d, each including four frames 506a, 506b, 506c, 506d. Similar to the previous embodiment, each frame series typically includes more than four frames, and the number of frame series is also greater than four. Furthermore, the number of frames per series of frames may be different from the number of series of frames.

  A horizontal parallax holographic stereogram consists of a holgel compressed along only one axis, whereas an omnidirectional parallax holographic stereogram is compressed along two axes, eg, as shown in FIG. Consists of a series of hologels. The two-dimensional array of predistortion images 508 is converted into holographic stereogram 510 holgels. In this example, the array consists of 225 predistortion images in the horizontal direction and 280 predistortion images in the vertical direction, ie the holographic stereogram is composed of an array of 225 × 280 holgels. The

  In one embodiment, average pixel values from the predistortion image may be used to generate a monochrome omnidirectional parallax holographic stereogram. Here, averaging is performed in two dimensions instead of one dimension. Therefore, the equation (2) shown above is replaced with the following equation (2a).

Here, num_horiz_pix represents the total number of pixels in the horizontal direction, num_vert_pix represents the total number of pixels in the vertical direction, and pAvg _{n, i} is 2 of the sum of the pixel values in row n and the sum of the pixel values in column i. The total of the two is further summed, and the average pixel value of the pixel (n, i) obtained by dividing the number of pixels in the horizontal direction (number of columns) and the number of pixels in the vertical direction (number of rows) is represented.

FIG. 18 shows a specific example of an omnidirectional parallax predistortion image 512 of 480 pixels × 640 pixels.
For any pixel n = 100, i = 80, the average pixel value pAvg is:

(Calculated for only one color, which is the green channel in this example.)
Different sets of reference data frames are used to generate the average pixel value pAvg vs. Br used for the horizontal parallax method. FIG. 19 shows an example of an appropriate set of reference data frames 514 for an omnidirectional parallax system. The set of reference data frames 514 includes a series of 10 frames (S1 to S10) each including 10 frames (F1 to F10). More than half of these reference data frames 514 are monochrome (i.e., not black, but are green, red, or blue), and the rest are black. The first frame series S1 does not include any black frames. In the second frame series S2, only the last frame is black. In the third frame series S3, the last two frames are black. In this manner, the black frames sequentially increase until the last frame series S10. In the frame S10, everything except the first frame is black.

  FIG. 20A shows a predistortion image 516 generated from the reference data frame set 514 of FIG. The predistortion images 516 are all the same. FIG. 20 (b) shows the uncompensated hologel 516 generated from the predistortion image 516. The same applies to all hologels. When physically reproduced, the predistortion image 516 becomes an “inaccurate” image 518.

  In the omnidirectional parallax print 520, the holographic stereogram generates an image corresponding to a change in the eye position 522 along the display plane 524, for example, along the axes x and y shown in FIG. When the observer views the display plane 524 at different positions along the diagonal 526, a different image 528 shown in FIG. Diagonal line 526 is the only part that includes a non-black area extending from the upper left corner of the print. The brightness of each different image 528 is measured and the Br value of that image is determined. In this case, each specific viewpoint along the diagonal line also corresponds to a specific pixel along the diagonal line of the predistortion image 516 (the predistortion images 516 are all the same, and here, which predistortion image 516 is the same). Is not important). The average pixel value pAvg of each pixel is determined, a set of points is determined, these points are plotted in the same manner as described above, and a curve is created by a set of these points. All predistortion images constitute a visible image, from which measured values are obtained. Therefore, only measurements along the diagonal of the actual stereogram print are sufficient. For example, in this case, when three points along the diagonal are measured, three images 528 of different brightness are obtained.

  FIG. 23 shows such a curve 530. Note that curve 530 is only part of the complete curve that is needed, and only covers a darker image (lower Br value) and a higher average pixel value pAvg.

  To obtain the remainder of the average pixel value pAvg vs. Br curve, a second set 534 of reference data frames is further generated, as shown in FIG. This second set of reference data frames 534 has a series of 10 frames (S1 to S10) each including 10 frames (F1 to F10). The first four frames of the first series S1, the first three frames of the second series S2, the first two frames of the third series S3, and the frames other than the first frame of the fourth series S4 Are all black. The first four frames of these non-black first frame series S1 are monochrome frames based on the color from which the curve is generated.

  A set of the same predistortion images 536 as shown in FIG. 25 (a) and a set of the same uncompensated hologel 538 as shown in FIG. 25 (b) are generated.

  Similarly to the case of creating the curve 530 shown in FIG. 23 described above, a set of Br values and corresponding average pixel values pAvg is generated. The points obtained from this second set of reference data frames 534 have higher brightness Br values and lower average pixel values pAvg, thus completing the reference curve 540 by plotting as shown in FIG. Can be made. In FIG. 26, a curved circled portion 542 is a portion obtained using the second set 534 of reference data frames.

  In a monochrome omnidirectional parallax holographic stereogram, the reference curve 540 can be used to derive an attenuation value that attenuates the average pixel value pAvg of the predistortion image pixels of the source data frame converted to the holographic stereogram. . Here, the above-described formula (1) and formulas (3) to (7) can be similarly used.

Omnidirectional parallax-color system Similar to the horizontal parallax-only color system described above, an omnidirectional parallax color system in which the holographic stereogram includes two or more color components of the three RGB color components, The predicted brightness Br of the RGB components is affected by the pixel values of the other RGB components. Again, the weighting factor Weight _X is used to compensate for the effect of the pixel values of other RGB components on the pixel values of a single RGB component. That is, Br can be expressed by the above equation (8). However, the curve used to obtain the weighting factor in a color system with only horizontal parallax is not ideal for use in an omnidirectional parallax system. Therefore, in this embodiment, an ideal new curve is generated.

  The method of generating the weighting coefficient curve in the embodiment of the omnidirectional parallax color system is the same as the method used for the color system with only the horizontal parallax described above with reference to FIGS. 12 and 13. Will be described with reference to FIGS. Here, generation of a green weight curve will be described as an example.

  FIG. 27A shows an array of source data frames. The array includes a series of 50 frames, each containing 100 frames. The frame series S1 to S25 are the same, and the frame series S26 to S50 are the same. In the series S1 to S25, the frames 1 to 50 are all the same, and have a green (G = 255) layer on a yellow (G = 255, R = 128, B = 0) layer. ˜100 are all the same, and are the same as frames 1 to 50 except that the left half of each frame is black. In the frame series S26 to S50, the frames 1 to 50 are the same as the frames 51 to 100 of the frame series S1 to S25, and the frames 51 to 100 are the same as the frames 1 to 50 of the frame series S1 to S25. It is.

  FIG. 27 (b) shows a hologel array obtained from the source data frame of FIG. 27 (a). The omnidirectional parallax print in this embodiment has 150 columns and 200 rows (eg, for a 60 × 45 mm print, each hogel is 0.3 mm square). Rows n1 to n100 are the same, and rows n101 to n200 are the same. In rows n1 to n100, columns i1 to i75 are the same, the hogel is divided into four parts, upper right and lower left black and upper left and lower right green, and columns i76 to i150 are the same and are all green. . In rows n101-n200, columns i1-i75 are the same, the hogel is divided into four parts, upper right and lower left black, and upper left and lower right yellow, columns i76-i150 are the same and are yellow. .

  When the omnidirectional parallax print is viewed again from the upper left corner on a display plane similar to the display plane 524 of FIGS. 21 and 22, an image as shown in FIG. 28 can be seen. The image 540 is divided into four, the upper left area is green, and the lower left area is yellow. The upper right region is also green, but it shows a darker color tone of green. The lower right region is also yellow, but this also shows a darker shade of yellow. The reference value Ref is obtained from the ratio between the upper right value and the upper left value. Also, a result r4 is obtained from the ratio between the lower right value and the lower left value.

FIG. 29 shows a reference green weighting factor curve plotted from nine points. The first point pt0 corresponds to pure green and is set to 1.0 on the Weight _G axis. The remaining eight points are derived from the omnidirectional parallax color holographic stereogram in a manner similar to that described above with respect to FIGS.

The value of Pt4 corresponding to red = 0.5, which is the midpoint on the red axis on the red-Weight _G plane, is r4 / Ref.

Pt1 is generated from a source data frame similar to FIG. 27 except that yellow has a red value R = 255 (rather than R = 128). The values r1 and Ref are derived from the image of the display plane of the omnidirectional parallax color holographic stereogram generated from these source frames. In the red-weight _G plane where red = 1.0, Pt1 = r1 / Ref.

Similarly, Pt2 is derived using a source data frame whose lower half includes R = 0, G = 255, B = 255. In the blue-weight _G plane where blue = 1.0, Pt2 = r2 / Ref.

  Similarly, Pt3 is derived using a source data frame whose lower half includes R = 255, G = 255, B = 255. At red = 1.0 and blue = 1.0, Pt3 = r3 / Ref.

Similarly, Pt5 is derived using a source data frame whose lower half includes R = 0, G = 255, B = 128. In the blue-weight _G plane where blue = 0.5, Pt5 = r5 / Ref.

  Similarly, Pt6 is derived using a source data frame whose lower half includes R = 128, G = 255, B = 128. At red = 0.5 and blue = 0.5, Pt6 = r6 / Ref.

  Similarly, Pt7 is derived using a source data frame whose lower half includes R = 128, G = 255, B = 255. At red = 0.5 and blue = 1.0, Pt7 = r7 / Ref.

  Similarly, Pt8 is derived using a source data frame whose lower half includes R = 255, G = 128, and B = 255. For red = 0.5 and blue = 0.5, Pt8 = r8 / Ref.

  For all these calculations, it is necessary to use the same value as the value Ref.

  Also, red and blue weighting coefficient graphs can be created in the same manner.

Also, the above-described equations (9) to (12) can be applied to an omnidirectional parallax color system. Further, when three weighting coefficient curves are created, each pixel value pallet _New compensated for the RGB components of the pixels of the predistortion image in the omnidirectional parallax color system is calculated using the above-described equations (13) to (21). can do.

FIG. 30 is a flow chart showing the steps necessary to perform pixel compensation when generating a color holographic stereogram. In steps S552, S554, and S556, a curve of Br vs. pAvg _ratio is derived for each of G, B, and R. In steps S558, S560, and S562, a weight factor curve is derived for each of G, B, and R. Once these reference curves are obtained, the pixels of the predistortion image of the source data frame that are converted to holographic stereograms can be compensated. In step S566, for each of R, G, and B, the average pixel value pAvg _{n, i} of each pixel of each predistortion image 564 is determined (in the case of horizontal parallax, the equation (2) is used, and in the case of omnidirectional parallax. , Using formula (2a)). Further, in step S568, a new pixel value palette of each pixel is determined from the related curve using the Br value and the weighting coefficient value (in FIG. 30, Y represents the current color, and X and Z represent other colors). Represents two colors).

  Further, the flowcharts of FIGS. 14 and 15 used for the holographic stereogram with only the horizontal parallax are applicable to the omnidirectional parallax holographic stereogram.

  The above-described processing may be realized only by hardware having a circuit designed specifically for necessary processing, or may be realized by a combination of hardware and software modules. A known computer system may be used.

  The module, particularly the module function, may be realized by either hardware or software. In software, a module is typically a process, program, or part thereof that performs a specific function or related functions. In hardware, a module is a functional hardware unit designed to be used with other components or modules. For example, the module may be implemented using independent electronic components, or may form part of an entire electronic circuit, for example, an application specific integrated circuit (ASIC). There are many other possibilities.

  FIG. 31 is a block diagram of a computer system 600 suitable for executing the processing described with reference to FIGS. The memory of the computer 602 is loaded with appropriate software for executing each step of the process for realizing the above-described technique. Using such a computer system 600, a program can be executed and a result can be obtained. This computer software is executed under a suitable operating system installed in computer system 600.

  The computer software can be interpreted by a processor, such as a central processing unit, and includes a set of programmed logic instructions that instruct the computer system 600 to perform a predetermined function specified by the instructions. The computer software may be a representation that includes a set of instructions recorded in any language, code, or notation, and these instructions are compatible either directly or after conversion to another language, code, or notation. Specific information processing system to execute a specific function.

  The computer software is programmed by a computer program that includes statements written in a suitable computer language. The computer program is processed using a complier to computer software having a binary format suitable for execution by the operating system. The computer software is programmed to include various software components or code that perform specific steps in the processes described above.

  The computer system 600 includes a computer 602, input / output devices such as a keyboard 604, a mouse 606, and an external memory device 608 (for example, one of a floppy disk drive, a CD drive, a DVD drive, and a USB flash memory drive). ), A display 610, a network interface connection terminal connected to the Internet 612, and an SLM 614 and a CCD 616 (or other mechanism for obtaining Br). The computer 602 includes a processor 622, a first memory such as a ROM 624, a second memory such as a RAM 626, a network interface 628 connected to an external network, and an input / output (I / O) An interface 630, a video interface 632 connected to a display, a storage device such as a hard disk 634, and a bus 636 are provided.

  The processor 622 executes an operating system and executes computer software under the operating system. Random access memory (RAM) 626, read-only memory (ROM) 624 and hard disk 634 are used under the control of processor 622.

  The video interface 632 is connected to the display 610 and supplies a video signal to be displayed on the display 610. User input for operating the computer 602 is provided from a keyboard 604 and a mouse 606.

  Here, the hard disk 634 is exemplarily shown as the internal storage device, but the computer 602 may include any other appropriate non-volatile storage medium.

  Each component of computer 602 is connected to a bus 636 that includes a data bus, an address bus, and a control bus so that the components can communicate with each other.

  Computer system 600 may be connected to one or more other similar computers via the Internet, a LAN, or other network.

  The computer software program can be provided as a computer program product. During normal use, the computer program can be stored on the hard disk 634. The computer software program may be recorded on a portable recording medium read by the external memory device 608, such as a CD-ROM. Alternatively, the computer software may be accessed directly from the network 612.

  In either case, the user can interact with computer system 600 using keyboard 604 and mouse 606 to operate programmed computer software running on computer 602.

  The source data frame is input to the processor 622 via the input / output interface 630 or the network interface 628. The processor 622 divides the source data frame and places the divided source data frame based on a mapping algorithm read into the system 600 from the computer program product and stored in the memory 626 to generate a plurality of predistortion images. .

  The compensation algorithm read into the system 600 from the computer program product is also stored in the memory 626. The processor 622 applies a compensation algorithm to the predistortion image to compensate for pixel saturation variations in the predistortion image that are associated with ghosting. When creating a weight and attenuation or lightness reference graph, a lightness value is input using the CCD 616.

  The compensated predistortion image is supplied to the SLM 614 via the input / output interface 628 and projected onto the holographic recording medium via the convergent lens to generate a holographic stereogram.

  Here, the computer system 600 is shown for explanation, but the above-described technology may be realized by using other configurations or types of computer systems. That is, computer system 600 is only one specific example of a particular type of computer system suitable for implementing the techniques described above.

  From the specification and specific examples of the present invention, those skilled in the art can easily conceive of other embodiments of the present invention. Furthermore, although specific terminology is used herein for the sake of clarity, it is not intended to limit the invention. The embodiments and preferred features of the invention described above are considered to be exemplary and the invention is defined by the claims and / or the description.

It is a figure explaining the process which produces | generates a holographic stereogram based on the prior art. It is a figure which shows the ghost in the image reproduced | regenerated from the holographic stereogram based on a prior art. FIG. 3A is a diagram showing a series of predistortion images based on the conventional technique. FIG. 3B is a diagram showing a conventional technique generated from a series of frames of the predistortion image shown in FIG. FIG. 3 (c) shows a series of frames of the holographic element of FIG. 3 (b) based on the prior art, compressed on a holographic recording medium. FIG. FIG. 4A is a diagram showing a predistortion image before applying a compensation algorithm based on the embodiment of the present invention, and FIG. 4B is a compensated predistortion image based on the embodiment of the present invention. FIG. 4C shows an uncompressed holographic element generated from the compensated predistortion image shown in FIG. 4B. It is a figure explaining the process which establishes the relationship between the average pixel value pAvg and Br about one of RGB components. It is a figure explaining calculation of the average pixel value pAvg of the pixel of a predistortion image. It is a figure explaining the process which establishes the relationship between average pixel value pAvg and Br. 8A is a diagram showing a graph of the average pixel value pAvg vs Br, FIG. 8B is a diagram showing a graph of the average pixel value pAvg _ratio vs Br, and FIG. It is a figure which shows the graph of brightness Br versus attenuation factor atten. It is a figure explaining calculation of the average pixel value pAvg of the pixel of a predistortion image. It is a figure which shows three curves of average pixel value pAvg _{ratio, G} vs. Br _G about different R and B values. FIG. 11A is a diagram illustrating a graph showing a relationship among the weighting coefficient Weight _G , the average pixel value pAvg _{ratio, R,} and the average pixel value pAvg _{ratio, B,} and FIG. FIG. 11C is a diagram illustrating a graph showing a relationship among the weighting coefficient Weight _R , the average pixel value pAvg _{ratio, G,} and the average pixel value pAvg _{ratio, B.} FIG. 11C illustrates the weighting coefficient Weight _B and the average It is a graph which shows the relationship between pixel value pAvg _{ratio, R} and average pixel value pAvg _{ratio, G.} 12A shows a set of source data frames, and FIG. 12B shows a source data frame reproduced from a holographic stereogram generated from the source data frame of FIG. 12A. It is a figure which shows one set of images. It is a figure explaining calculation of the value which plots the graph of Fig.11 (a) from the image of FIG.12 (b). It is a flowchart of the process which produces | generates a holographic stereogram. It is a flowchart of the process which compensates the dispersion | variation in the saturation degree of a predistortion image. It is a figure which shows the specific example of a set of data source frames for producing | generating an omnidirectional parallax holographic stereogram. FIG. 6 is a schematic diagram of an array of predistortion images and a holgel for generating an omnidirectional parallax holographic stereogram. It is a figure which shows the specific example of the predistortion image of an omnidirectional parallax. FIG. 6 shows a first set of reference data frames for generating an omnidirectional parallax holographic stereogram. 20A is a diagram illustrating a predistortion image generated from the first set of reference data frames in FIG. 19, and FIG. 20B is a diagram illustrating the first set of reference data frames in FIG. FIG. 3 shows the uncompensated hologel produced. FIG. 6 is a perspective view of a display plane and an omnidirectional parallax holographic stereogram print. It is a top view of the display plane of an omnidirectional parallax holographic stereogram print. FIG. 20 shows a graph generated from the first set of reference data frames of FIG. 19. FIG. 6 shows a second set of reference data frames for generating an omnidirectional parallax holographic stereogram. FIG. 25A is a diagram illustrating a predistortion image generated from the second set of reference data frames in FIG. 24, and FIG. 25B is a diagram illustrating the second set of reference data frames in FIG. FIG. 3 shows the generated uncompensated hologel. FIG. 25 illustrates a combined graph generated from the first set of reference data frames of FIG. 19 and the second set of reference data frames of FIG. 24. FIG. 27A is a diagram showing an array of source data frames for generating a reference curve in a color omnidirectional parallax system, and FIG. 27B is an array of hologels generated from the source data frames of FIG. FIG. FIG. 28 is an image from a holographic stereogram print having the holgel of FIG. It is a figure which shows the reference | standard green weighting coefficient curve plotted using the value from the image of FIG. It is a flowchart of the process which produces | generates a color omnidirectional parallax holographic stereogram. It is a figure which shows the structure of the computer system which produces | generates a holographic stereogram from several source data frames.

Claims (9)

In an image processing method for generating a holographic stereogram from a plurality of source data frames,
Dividing the plurality of source data frames into a plurality of sections;
Mapping the plurality of sections to generate a plurality of predistortion images;
Compensating for variations in saturation of the plurality of predistortion images;
Compressing the plurality of predistortion images to generate a plurality of holographic elements;
Recording the plurality of holographic elements on a holographic recording medium to generate the holographic stereogram.   The image processing method according to claim 1, further comprising a step of compensating the plurality of predistortion images by applying a compensation algorithm to one or more pixels of the plurality of predistortion images.   The compensation algorithm adjusts one or more pixels of the plurality of predistortion images when the plurality of holographic elements are recorded on the holographic recording medium to at least partially compensate for subsequent saturation variations. The image processing method according to claim 2, wherein compensation is performed.   4. The image processing method according to claim 2, wherein one or more attenuation factors are applied to one or more pixels of the plurality of predistortion images.   The attenuation rate applied to the pixel reduces the pixel value to the lowest expected level of non-zero pixel values of a plurality of holographic elements recorded on the holographic recording medium. 5. The image processing method according to 4.   Establishing a reference relationship between the average pixel value of the plurality of pixels of the plurality of predistortion images of the second holographic stereogram and the brightness of the image reproduced from the second holographic stereogram; 6. The image processing method according to claim 1, further comprising:   7. The image processing method according to claim 6, wherein the reference relationship is determined from a series of reference source data frames having a color gradation. The above standard relationship is
(A) generating a series of reference source data frames having a color gradation ;
(B) segmenting and mapping the series of reference source data frames to generate a set of predistortion images;
(C) compressing the set of predistortion images to generate a set of holographic elements;
(D) recording the set of holographic elements on a second holographic recording medium to generate a second holographic stereogram;
(E) measuring the brightness of the image reproduced from the second holographic stereogram;
(F) plotting a graph of average pixel values of the pixels of the set of predistortion images against the brightness of the image reproduced from the second holographic stereogram. The image processing method according to claim 6.   9. The image processing method according to claim 8, wherein the second holographic recording medium is the same type of medium as the holographic recording medium.

Images